https://www.datalab.to/blog Product releases, benchmarks, and case studies from Datalab — document intelligence APIs for conversion, OCR, and structured extraction. en https://www.datalab.to/blog/document-agent-beta https://www.datalab.to/blog/document-agent-beta Mon, 24 Aug 2026 00:00:00 GMT Datalab Document Agent is now in closed beta: specify what you want, provide sample documents, iterate on what correct means, and our agent will devise a custom suite of verifiers and tools built on our state of the art models and pipelines to produce the right document workflow for your needs. https://www.datalab.to/blog/form-filling-on-the-document-agent https://www.datalab.to/blog/form-filling-on-the-document-agent Mon, 24 Aug 2026 00:00:00 GMT Form filling now runs on our document agent: geometry is measured rather than guessed, and every fill is checked against the page it produced. Building the benchmark to prove it also turned up three defects in the version we had been shipping, all of them silent. https://www.datalab.to/blog/segment-pricing https://www.datalab.to/blog/segment-pricing Wed, 19 Aug 2026 00:00:00 GMT We cut the price of document segmentation by 92% — and for the most common use, by 95%. https://www.datalab.to/blog/wcag-accessible-pdfs https://www.datalab.to/blog/wcag-accessible-pdfs Wed, 19 Aug 2026 00:00:00 GMT A new processor that turns a PDF — including a scan that is nothing but pixels — into a tagged PDF built for assistive technology, checked against a battery of automated tests and a real screen reader. https://www.datalab.to/blog/extractbench-scoring-bug https://www.datalab.to/blog/extractbench-scoring-bug Thu, 13 Aug 2026 00:00:00 GMT The benchmark shipped with major scoring bugs. After fixing them, we score 93.6% on the benchmark, leading all other contenders. We encourage you to run your own evals, and not just trust vendor benchmarks. https://www.datalab.to/blog/pdf-to-jats-xml https://www.datalab.to/blog/pdf-to-jats-xml Thu, 13 Aug 2026 00:00:00 GMT A new processor that turns raw scientific paper PDFs into DTD-validated JATS 1.2 XML, with every run carrying a conformance verdict you can audit. https://www.datalab.to/blog/apparelwerks-datalab-casestudy https://www.datalab.to/blog/apparelwerks-datalab-casestudy Mon, 10 Aug 2026 00:00:00 GMT How a made-to-measure clothing studio used Datalab to build a production system where every inch counts. https://www.datalab.to/blog/word-bboxes-degradation-density-rotation https://www.datalab.to/blog/word-bboxes-degradation-density-rotation Tue, 04 Aug 2026 00:00:00 GMT Bounding boxes for bad scans, dense pages, and sideways text. https://www.datalab.to/blog/high-accuracy-update https://www.datalab.to/blog/high-accuracy-update Fri, 24 Jul 2026 00:00:00 GMT High accuracy mode is now more accurate while making fewer changes to the page, powered by new verification models trained on-distribution with Chandra. https://www.datalab.to/blog/marker-2 https://www.datalab.to/blog/marker-2 Mon, 20 Jul 2026 00:00:00 GMT Marker 2 is a rewrite of our open-source PDF-to-markdown converter. It runs on CPU, picks a speed/accuracy mode automatically, and beats comparable pipeline OCR systems on both accuracy and throughput on olmOCR-bench. https://www.datalab.to/blog/kin-datalab-casestudy https://www.datalab.to/blog/kin-datalab-casestudy Fri, 17 Jul 2026 00:00:00 GMT How Kin rebuilt the way adjusters cite insurance policies, achieving compliance-grade accuracy with Chandra and custom processors. https://www.datalab.to/blog/trusting-vendor-benchmarks https://www.datalab.to/blog/trusting-vendor-benchmarks Tue, 14 Jul 2026 00:00:00 GMT A competitor commissioned a long-table extraction benchmark, and scored Datalab at 34% recall. We rebuilt our extraction and now have the top recall (99.1%) and lead on precision too. Still, you should take all vendor benchmarks with a grain of salt and run your own evals. https://www.datalab.to/blog/learnboost-datalab-casestudy https://www.datalab.to/blog/learnboost-datalab-casestudy Wed, 08 Jul 2026 00:00:00 GMT How an AI-powered learning platform boosted retention by replacing its document conversion with Marker. https://www.datalab.to/blog/custom-processors https://www.datalab.to/blog/custom-processors Tue, 07 Jul 2026 00:00:00 GMT Custom Processors turn plain-language instructions and a few examples into a processor that parses your documents exactly the way you need. https://www.datalab.to/blog/anti-hallucination https://www.datalab.to/blog/anti-hallucination Tue, 30 Jun 2026 00:00:00 GMT Hallucinations are a fact of how LLMs work. For mission-critical documents, even one is too many. Here's the defense-in-depth system we use to drive them toward zero. https://www.datalab.to/blog/word-bounding-boxes-and-confidence https://www.datalab.to/blog/word-bounding-boxes-and-confidence Tue, 23 Jun 2026 00:00:00 GMT Making OCR output auditable. https://www.datalab.to/blog/free-to-start-pay-as-you-go https://www.datalab.to/blog/free-to-start-pay-as-you-go Thu, 18 Jun 2026 00:00:00 GMT Datalab now has a free tier and pay-as-you-go pricing. Start parsing for free with a monthly usage allowance, then pay only for the pages you actually process — no subscription, no minimum, no plan to pick. https://www.datalab.to/blog/introducing-lift https://www.datalab.to/blog/introducing-lift Thu, 18 Jun 2026 00:00:00 GMT lift is a 9B open-weights vision model that extracts structured JSON from PDFs and images by passing a schema. It scores 90.2% field accuracy on our 225-document extraction benchmark - the best small self-hostable model we've tested. https://www.datalab.to/blog/chandra-2.1-release https://www.datalab.to/blog/chandra-2.1-release Wed, 17 Jun 2026 00:00:00 GMT Announcing Chandra 2.1, a smaller, faster model that improves on multilingual and table accuracy. https://www.datalab.to/blog/turbo-extraction https://www.datalab.to/blog/turbo-extraction Thu, 11 Jun 2026 00:00:00 GMT Turbo mode for the Datalab extraction API: structured JSON from any document at a median of ~12 seconds per document. https://www.datalab.to/blog/datalab-nevada-county-historical-archive-case-study https://www.datalab.to/blog/datalab-nevada-county-historical-archive-case-study Fri, 05 Jun 2026 00:00:00 GMT How the Nevada County Historical Archive uses Datalab's Chandra model to transcribe 200K+ pages of handwritten Gold Rush-era records 150X faster, making a century of California history searchable. https://www.datalab.to/blog/introducing-balanced-extraction-mode https://www.datalab.to/blog/introducing-balanced-extraction-mode Thu, 04 Jun 2026 00:00:00 GMT Balanced is our new default extraction mode: higher accuracy, with reasoning and independent verification baked into every response. It knows when a value isn't in the document instead of making one up. https://www.datalab.to/blog/eu-data-residency https://www.datalab.to/blog/eu-data-residency Fri, 29 May 2026 00:00:00 GMT Datalab now offers data residency controls, starting with a new processing location in the Netherlands. Keep your document content and processing entirely within the EU, using the same API you already use. https://www.datalab.to/blog/surya-2 https://www.datalab.to/blog/surya-2 Wed, 27 May 2026 00:00:00 GMT Surya OCR 2 is a 650M-parameter open-source OCR model that scores 83.3% on olmOCR-bench, hits 87.2% on a 91-language multilingual eval, and runs on CPU, GPU, and MPS. https://www.datalab.to/blog/managed-batch-processing-is-now-live https://www.datalab.to/blog/managed-batch-processing-is-now-live Wed, 29 Apr 2026 00:00:00 GMT Introducing Datalab's Managed Batch Processing — give us a bucket, we handle the rest. Process millions of pages without managing GPUs, scaling infrastructure, or orchestrating workloads. https://www.datalab.to/blog/datalab-radical-ai-case-study https://www.datalab.to/blog/datalab-radical-ai-case-study Thu, 26 Mar 2026 00:00:00 GMT Radical AI ($55M Seed+) ingests millions of pages of dense academic literature through Datalab to feed the autonomous-discovery systems shrinking materials R&D from decades to weeks. https://www.datalab.to/blog/chandra-2 https://www.datalab.to/blog/chandra-2 Wed, 18 Mar 2026 00:00:00 GMT Chandra 2 is a 4B parameter OCR model with state of the art benchmarks, layout blocks with bounding boxes, and structured output for diagrams and charts. https://www.datalab.to/blog/chandra-1-5 https://www.datalab.to/blog/chandra-1-5 Thu, 22 Jan 2026 00:00:00 GMT Announcing Chandra 1.5 with dramatically improved table extraction, chemistry support with SMILES output, diagram rendering, and significant latency improvements. https://www.datalab.to/blog/datalab-rely-case-study https://www.datalab.to/blog/datalab-rely-case-study Fri, 16 Jan 2026 00:00:00 GMT Rely audits commercial real-estate transactions worth millions in misvalued assets. Running Datalab on-prem at 12.5 pages/sec lets them ingest 500,000-page data rooms before deals close — without resident documents ever leaving their network. https://www.datalab.to/blog/cofactr-datalab-casestudy https://www.datalab.to/blog/cofactr-datalab-casestudy Tue, 06 Jan 2026 00:00:00 GMT Cofactr (YC W'22) serves regulated hardware manufacturers — Amazon Robotics among them. Datalab parses inconsistent BOMs, quotes, and spec sheets across 5,000+ suppliers with the fidelity their compliance and traceability work demands. https://www.datalab.to/blog/extracting-hyperlinks-from-pdfs https://www.datalab.to/blog/extracting-hyperlinks-from-pdfs Fri, 19 Dec 2025 00:00:00 GMT Learn how to extract hyperlinks from PDFs using the Datalab API, preserving both link text and destinations. https://www.datalab.to/blog/automatically-fill-pdf-forms-with-ai https://www.datalab.to/blog/automatically-fill-pdf-forms-with-ai Wed, 17 Dec 2025 00:00:00 GMT Automatically fill PDF forms using AI with our new form filling feature. https://www.datalab.to/blog/introducing-forge-evals https://www.datalab.to/blog/introducing-forge-evals Wed, 10 Dec 2025 00:00:00 GMT A new tool to evaluate how various settings and modes impact document parsing quality across PDFs, DOCX files, and spreadsheets. https://www.datalab.to/blog/datalab-benchmarks-evals https://www.datalab.to/blog/datalab-benchmarks-evals Tue, 09 Dec 2025 00:00:00 GMT View scores and compare model output across a sample of diverse pages, or get evals on your own documents. https://www.datalab.to/blog/launch-week-faster-tracked-changes-outputs https://www.datalab.to/blog/launch-week-faster-tracked-changes-outputs Fri, 05 Dec 2025 00:00:00 GMT Performance improvements for Tracked Changes output parsing, plus native spreadsheet support in Forge. https://www.datalab.to/blog/launch-week-spreadsheet-parsing https://www.datalab.to/blog/launch-week-spreadsheet-parsing Thu, 04 Dec 2025 00:00:00 GMT Native spreadsheet support in the Datalab API for accurate table segmentation and parsing. https://www.datalab.to/blog/launch-week-introducing-agni-solving-multi-page-section-hierarchy-in-ocr https://www.datalab.to/blog/launch-week-introducing-agni-solving-multi-page-section-hierarchy-in-ocr Wed, 03 Dec 2025 00:00:00 GMT Introducing Agni, a new model that maintains consistent section hierarchy across multi-page documents. https://www.datalab.to/blog/launch-week-chandra-is-faster-again https://www.datalab.to/blog/launch-week-chandra-is-faster-again Tue, 02 Dec 2025 00:00:00 GMT Announcing Chandra Small, a latency-optimized model that achieves 2-3x faster speeds with minimal performance degradation. https://www.datalab.to/blog/launch-week-chandra-1-1 https://www.datalab.to/blog/launch-week-chandra-1-1 Mon, 01 Dec 2025 00:00:00 GMT Announcing Chandra 1.1 with improved layout, math, tables, and multilingual performance. https://www.datalab.to/blog/speeding-up-chandra https://www.datalab.to/blog/speeding-up-chandra Mon, 17 Nov 2025 00:00:00 GMT How we achieved 3x faster Chandra performance using Eagle3 speculative decoding without sacrificing accuracy. https://www.datalab.to/blog/saturating-the-olmocr-benchmark https://www.datalab.to/blog/saturating-the-olmocr-benchmark Thu, 13 Nov 2025 00:00:00 GMT Our analysis reveals fundamental limitations in the olmOCR benchmark and why traditional edit-distance benchmarks provide diminishing returns for OCR evaluation. https://www.datalab.to/blog/extract-tracked-changes-metadata https://www.datalab.to/blog/extract-tracked-changes-metadata Wed, 12 Nov 2025 00:00:00 GMT New feature enabling extraction of tracked changes metadata from Word documents for contract review workflows. https://www.datalab.to/blog/high-fidelity-ocr-drives-accurate-structured-extraction https://www.datalab.to/blog/high-fidelity-ocr-drives-accurate-structured-extraction Tue, 04 Nov 2025 00:00:00 GMT How superior OCR technology enhances the accuracy of structured data extraction from documents. https://www.datalab.to/blog/introducing-chandra https://www.datalab.to/blog/introducing-chandra Thu, 30 Oct 2025 00:00:00 GMT Meet Chandra, our latest OCR model that tops independent benchmarks and brings powerful new capabilities to document processing. https://www.datalab.to/blog/view-your-api-requests-in-our-playground https://www.datalab.to/blog/view-your-api-requests-in-our-playground Tue, 07 Oct 2025 00:00:00 GMT Visualize your API requests directly within the Playground to audit and debug your document processing. https://www.datalab.to/blog/launch-week-playground-examples https://www.datalab.to/blog/launch-week-playground-examples Tue, 30 Sep 2025 00:00:00 GMT We've added practical examples to the Playground that you can explore directly. https://www.datalab.to/blog/launch-week-high-accuracy-mode https://www.datalab.to/blog/launch-week-high-accuracy-mode Fri, 26 Sep 2025 00:00:00 GMT Introducing High Accuracy Mode, which combines our proprietary models with frontier LLMs to handle challenging documents. https://www.datalab.to/blog/launch-week-layout-model https://www.datalab.to/blog/launch-week-layout-model Thu, 25 Sep 2025 00:00:00 GMT Significant upgrades to our OCR and Layout models, achieving state-of-the-art performance on olmOCR-bench. https://www.datalab.to/blog/launch-week-document-segmentation https://www.datalab.to/blog/launch-week-document-segmentation Wed, 24 Sep 2025 00:00:00 GMT New Document Segmentation feature that automatically detects page boundaries within multi-document PDF files. https://www.datalab.to/blog/launch-week-new-playground https://www.datalab.to/blog/launch-week-new-playground Tue, 23 Sep 2025 00:00:00 GMT Unveiling our updated Playground with parsing, refinement, segmentation, and extraction capabilities. https://www.datalab.to/blog/cracking-math-ocr https://www.datalab.to/blog/cracking-math-ocr Fri, 19 Sep 2025 00:00:00 GMT How high-quality mathematical text extraction enables training of reasoning-capable language models. https://www.datalab.to/blog/structured-extraction-with-datalab-api-and-handling-long-documents https://www.datalab.to/blog/structured-extraction-with-datalab-api-and-handling-long-documents Mon, 08 Sep 2025 00:00:00 GMT How to use Datalab's Marker API for structured extraction from PDFs using JSON schemas, plus strategies for processing lengthy documents. https://www.datalab.to/blog/structured-extraction-citations https://www.datalab.to/blog/structured-extraction-citations Tue, 19 Aug 2025 00:00:00 GMT Structured Extraction now includes citation support, enabling users to verify information sources within documents. https://www.datalab.to/blog/parse-pdfs-just-the-way-you-want https://www.datalab.to/blog/parse-pdfs-just-the-way-you-want Thu, 14 Aug 2025 00:00:00 GMT Introducing Marker Prompt API and Forge Parse for customizing PDF parsing outputs through prompts. https://www.datalab.to/blog/releases-forge-extract https://www.datalab.to/blog/releases-forge-extract Thu, 07 Aug 2025 00:00:00 GMT Extract structured data from PDFs using custom schemas with our new Forge Extract feature. https://www.datalab.to/blog/datalab-purchaser-case-study https://www.datalab.to/blog/datalab-purchaser-case-study Wed, 16 Jul 2025 00:00:00 GMT Manufacturing procurement runs on messy quotes and thousand-row BOMs. Datalab handles the inconsistency so Purchaser.ai's team ships faster purchasing decisions — and keeps the customers who'd otherwise churn on bad parses. https://www.datalab.to/blog/releases-sdk https://www.datalab.to/blog/releases-sdk Wed, 09 Jul 2025 00:00:00 GMT Announcing the official Datalab SDK for Python with simplified APIs and powerful features. https://www.datalab.to/blog/datalab-revisiondojo-case-study https://www.datalab.to/blog/datalab-revisiondojo-case-study Tue, 01 Jul 2025 00:00:00 GMT RevisionDojo's edtech platform serves 310K+ students who need study materials parsed on demand. Retiring their self-hosted screenshot pipeline for Datalab cut processing cost 10x and brought average parse time to ~10 seconds. https://www.datalab.to/blog/datalab-gamma-case-study https://www.datalab.to/blog/datalab-gamma-case-study Tue, 17 Jun 2025 00:00:00 GMT Gamma's 50M users turn PDFs into working presentations. Switching parsers to Datalab lifted PDF-import conversion 40% — over a million PDFs a week now flow through the engine without dropping layout or table fidelity. https://www.datalab.to/blog/building-better-document-intelligence-for-an-ai-first-world https://www.datalab.to/blog/building-better-document-intelligence-for-an-ai-first-world Thu, 12 Jun 2025 00:00:00 GMT Why highly accurate document intelligence matters more than ever in an AI-first world.